IBM Rochester

3605 Highway 52 North, Rochester, MN, USA 55901

Introduction  life after Baldrige

IBM Rochester won its Baldrige award in 1990, at a time when the parent corporation
was in some trouble. Early critics of the value of the Baldrige award frequently (and of
course, prematurely) cited this apparently bizarre outcome as yet another nail in the
award’s coffin. For IBM’s view of this period, see the exhibit IBM’s corporate history,
a summary of material from IBM’s corporate web site. For a contemporary view, see
IBM today

IBM today is not the organization that it was in the late 80's. IBM Rochester is no
exception and is a much changed organization, having built on its solid foundation of
total quality management employed in the late 80's to become the dominant, world
class, global supplier of mid-range commercial computer solutions. IBM Rochester’s
story is interesting in the context of the Baldrigeplus case study collection because of
the eight, it represents the oldest award − ten years have passed since a 17-person
IBM team put together the 1990 winning application.

It’s also interesting because it still reads well − it could be a 1999 application
(although not at the national Malcolm Baldrige level, where standards are now much
higher − maybe for a state award?). That’s an important observation to me (read
through the material and see if you agree) because as an enthusiast, I argue that
Baldrige is not just another passing management fad, not just another tool du jour, but
rather is a continually adjusted, ‘at the validated leading edge’ (Harry Hertz, personal
communication, 1999) approach to management excellence.

The case study begins with a run through the 1990 application material, represented
by, extracted and edited from, the 1991 second edition of the 20 page Baldrige
application summary and the two-page fact sheet that NIST publishes.

For a rounded picture check the IBM Rochester exhibits which deal with specific
issues; the quality dashboard; customer feedback and, customer satisfaction and
market share exhibits.
Malcolm Macpherson

1
Introduction to IBM Rochester
The IBM site in Rochester, MN is a 1 mile long facility consisting of 32 interconnected
buildings covering 32 acres, with 7,000 employees. It is the ‘home of IBM's AS/400
commercial mid-range computer,’ where both development and manufacturing are
performed. It is also the location of manufacture for the RS/6000, IBM's mid-range
commercial UNIX-based computer system. The AS/400 and RS/6000 products span
the range from desktop PC's to mainframes. There’s also some disk substrate
manufacturing at this site. Most of the Rochester site staff are engineers, programmers
and professionals. At the time of writing (early 1999) one AS/400 was shipped every
12 minutes, and the AS/400 business was worth about 15% of IBM’s total revenue.

The AS/400 is the world’s most popular multi-user business computing system, with:
• a broad range of customers
• 600,000 systems installed in 150 countries
• 97% of FORTUNE 100 industrials are users
• 28,000 commercial applications in use (more than any competitor) written by
IBM’s business partners and customers, in 40 national languages
• Y2K certified

The RS/6000, ‘a powerful, tailorable, UNIX-based system,’ has:

• an installed base of 800,000 in 150 countries
• 15,000 software applications
• 12,000 plus solution developers
• Y2K ready
• Guaranteed open
• 82% of all Internet servers
• the F50 Internet Powersolution won Info World 1997/98 product of the year
• aka ‘Deep Blue’ (the chess playing computer)
• the Mars Lunar Module, and the US Department of Energy are high-profile users.

Both systems are developed and manufactured world-wide, and both products and
their manufacturing sites are world-wide ISO 9001 and 14001 registered.

IBM Rochester in 1990 (From NIST’s Baldrige fact sheet)

Caution Some information from the 1991 Baldrige fact sheet utilised in this case study
may not reflect current processes or practices employed by IBM Rochester.

In 1990 the concept of quality at IBM Rochester was linked directly to the customer.
Detailed features were crafted by analyzing the needs and expectations of existing
and potential owners of the firm’s computer hardware and. At every step, customers
were directly involved in product from design to delivery − through advisory councils,
global information systems, trials of prototypes, and numerous other feedback
mechanisms.

The IBM Rochester quality culture had been transformed from reliance on technology-
driven processes delivering products to market-driven processes directly involving
suppliers, business partners, and customers delivering solutions. A 30% improvement
in productivity occurred between 1986 and 1989. Product-development time for new
mid-range compute systems had been reduced by more than half, while the
manufacturing cycle had been trimmed 60% since 1983. Customers benefited from a
3-fold increase in product reliability; an increase from 3 to 12 months in the product
warranty period; and a cost of ownership that was among the lowest in the industry.
IBM’s share of the world market for intermediate computers increased in both 1988
and 1989.
2
IBM Rochester at a glance  in 1990
Rochester manufactures intermediate computer systems − currently the AS/Entry
Systems and the Application System/400 (AS/400). More than 400,000 IBM Rochester
systems had been installed worldwide by 1990.

The Rochester location also made hard disk drives, sales of which accounted for
about a fifth of IBM Rochester's revenues in 1989.

IBM Rochester provided employment to more than 8,100 people and was responsible
for product development and US manufacturing. In addition, its processes were
implemented in plants in Japan, Mexico, the UK and Italy.

Rochester excellence  customer satisfaction

IBM Rochester had recently strengthened its strategic quality initiatives by formulating
improvement plans based on six critical success factors; improved product and
service requirements definition, an enhanced product strategy, a six-sigma defect
elimination strategy, further cycle time reductions, improved education, and
increased employee involvement and ownership.

Each senior manager “owned” one of the six factors and assumed responsibility for
plans and implementation. Progress toward achieving improvement goals was closely
monitored. Support processes were a part of this network.

Continuous improvement of support processes at IBM Rochester rested on aggressive

world-wide benchmarking, a process that analyses products and services to
determine the best of the breed in all industries. Over 350 teams were in place.
Scores of benchmarking studies had been completed.

Quality goals were established in five-year business plans and annual operations
plans. Strategic targets were derived from comprehensive benchmarking processes.
With the aid of financial planning models and mathematical decision-making tools,
quality priorities were set, and the resources human and capital necessary to carry
out these priorities were determined.

Most plans for achieving quality objectives originated with employees, and cross-
functional teams identified needs for equipment, staffing, education, and process
development. Each quality improvement plan had an owner, a managerial or non-
managerial employee who headed the project team. With employee input, objectives
and requirements were established for each employee, and a measurement system for
monitoring progress was developed in advance of the project. Close coordination and
efficient communication were ensured through regular planning meetings, in which
key suppliers and customers participates.

IBM Rochester invested heavily in education and training, the equivalent of 5% of its
payroll. Employees, supported by IBM's tradition of full employment, were
encouraged to develop the skills and expertise for a variety of jobs. In 1989, about a
third of the work force moved into new positions, and 13% were promoted. IBM
Rochester was implementing a Management System for Education, which offered skill
planning, needs assessment, individual education plans, and educational road maps
on-line. Job flexibility and security, ample opportunity for advancement, and a well-
developed recognition process were among factors contributing to rates of
absenteeism and turnover well below national averages. Employee contributions to
quality improvement were recognised in a variety of ways, including luncheons,
receptions, and monetary and non-monetary awards.

3
Morale was high, as determined in IBM annual surveys, and by an independently
conducted survey that compared levels of worker satisfaction at 34 US companies.

Equipping workers with the tools and information they need to accomplish quality and
customer satisfaction objectives was also a priority. Over 11,000 on-line terminals
provided employees with world-wide access to extensive communication
capabilities, databases, and design and analysis tools. For example, since 1966, IBM
had invested more than $300m in improving its processes and information systems.
Such investments, many of them designed to improve problem-solving capabilities
ensuring that defects are prevented rather than detected after they occur; paid for
themselves. Capital spending on equipment for defect detection declined 75% during
the 1980s, and write-offs as a proportion of manufacturing output dropped 55%.

At the start of the product planning process suppliers were included as partners to
ensure that new hardware and software achieved IBM’s goals for manufacturability,
serviceability, reliability, performance, and cost. Accounting for about 30% of
production output, IBM Rochester's approximately 700 production suppliers were
expected to provide defect-free shipments- and to keep pace with progress in
improving quality and reducing development and manufacturing cycles. Suppliers
were trained, audited, and certified, and required to submit quality plans. IBM
Rochester shared its own state-of-the-art technology with suppliers and, in turn,
suppliers provide valuable expertise to IBM. Since 1984, IBM Rochester employees
had instructed more than 1,909 supplier employees on continuous flow
manufacturing, statistical process control, and design of experiments.

The Rochester quality process was a continuous loop that began, ended, and began
again with the customer. Of the approximately 40 data sources analysed to guide
improvement efforts, most either provided information on customers' product and
service requirements or guide steps to refine these expectations into detailed
specifications for new IBM offerings. Customers were also active participants. For
example, customers and business partners- representing over 4,500 businesses
worldwide participated on customer advisory councils throughout the development of
the AS/400.

To strengthen its competitive quality position, IBM Rochester was aiming for a tenfold
improvement in key quality areas by I 991, a hundredfold improvement by 1993, and
a six sigma level of defects by 1994.

IBM Rochester in 1990  from the Baldrige Application Summary

In its 1990 Malcolm Baldrige National Quality Award application summary, IBM’s
application team described their part of the IBM organisation as “a leader in market-
driven quality planning and implementation within the IBM Corporation.”

The Rochester site at that time reported to two of the seven lines of business in the
IBM Corporation  the Application Business Systems (ABS) line of business for its
AS/4000 family of products, and to the Enterprise Systems (ES) line of business for its
hard disk drive family of products. Inside the US sales and service support was
provided by the US Marketing and Service line of business, and outside the US this
line of business was supported by the World Trade Organisations. Corporate quality
guidance to all lines of business was the responsibility of the ‘Market-Driven Quality
Organisation.’

4
 The IBM organisation, March 1991
IBM corporation
management
committee

IBM World Trade IBM IBM United Sates

Europe, Middle Market-driven
East, Africa quality
Asia-Pacific Group

Canada/Latin
America

Application Enterprise
business systems systems line Other lines of business,
line of business of business marketing and service

Personal systems

Technology products
Comms systems

IBM San Jose Programming

IBM Rochester systems
IBM Poughkeepsie

IBM Endicott
Other IBM sites

Source IBM Rochester’s Malcolm Baldrige National Quality Award

summary publication, second edition, IBM, Rochester, March 1991

IBM Rochester had worldwide development and US manufacturing responsibility for

the AS/400 computer system and hard disk storage devices, with sister manufacturing
facilities in Guadalajara, Mexico; Sumare, Brazil; Havant, England; Santa Palomba,
Italy; and Fujisawa, Japan. IBM Rochester, as well as over 20 other IBM locations,
provided programming development and support for the AS/400 system.

At that time more than 400,000 AS/400 and predecessor systems (System/38 and
System/36) have been installed worldwide. The AS/400 system software had been
translated into 28 national language versions to support the over 60% of AS/400
systems installed outside the US.

The AS/400 system was sold directly by IBM as well as by over 7,000 independent
business partners worldwide. A 6,000-member independent user's group called
COMMON provided valuable input to IBM Rochester for future product needs and
requirements.

The storage devices produced in 1990 at Rochester included hard disk drives for the
AS/400 and RS/6000 systems, and the PS/2 computer. Rochester also provided
hard disk drives to other computer equipment manufacturers. These hard disk drives
were considered leaders in their class in reliability, performance and diagnostic capa-
bilities.

The Rochester site housed electronic hardware and software systems development,
hard disk drive development and manufacturing, as well as prototype facilities for
machining, plastic molding, heat treating, plating and integrated circuit processing.
IBM Rochester also had its own tool and model makers who built test and process
equipment.

5
Sixty-six percent of IBM Rochester's then 8,100-person workforce was comprised of
engineers, programmers and other professionals, who continuously recorded very
high morale − the highest among IBM sites − as measured by annual opinion surveys.
The site’s safety record was 60% better than the industry average, and its turnover
rate was significantly below the industry average.

Quality deployment
The cornerstones of the IBM culture, said the 1990 application summary, were IBM’s
basic beliefs, respect for the individual, the best customer service, and pursuit of
excellence − fundamental cornerstones of the business that had remained unchanged
for 75 years.

Corporate Policy Letters and Corporate Instructions (CIs) provided guidance for
implementing these basic beliefs. For example, CI 101 required that all business
processes, product and non-product, improve continually to make them more
efficient, effective, and adaptable.

CI 105 stated that each new product must be superior in quality to previous products
(both IBM's and competitors'). Before announcing a product, each IBM development
lab had to demonstrate to the corporate quality organization and other line executives
that it was complying with CI 105.

Goals and market-driven quality principles were developed from the Corporate Policy
Letters and Instructions. IBM's Market-Driven Quality Principles were:
• make the customer the final arbiter
• understand our markets
• commit to leadership in the markets we choose to serve,
• deliver excellence in execution across our enterprise.

IBM Rochester had a quality policy, a vision, and a set of quality goals that are
deployed throughout the site to individual departments. This policy was supported by
the site vision:
• Customer − the final arbiter
• Products and services − first with the best
• Quality − excellence in execution
• People − enabled, empowered, excited, and rewarded.

Quality objectives were integrated into employee performance plans and agreed
upon by both the manager and the employee. This process fostered employee
participation and empowerment. Each element of quality improvement built a solid
foundation for achieving Rochester's business objectives.

The illustration below (next page) is the 1999 version of Rochester’s quality journey.
In 1990 the staircase ended at 1990-1994, and the mid-decade goal was “undisputed
leadership in customer satisfaction. IBM Rochester is now strategically positioned to
implement an expanded set of initiatives that build upon prior successes.

“The new initiatives consisted of six critical success factors and proactive customer
satisfaction processes that integrate all elements of development, manufacturing,
marketing and service. The goal of IBM Rochester is to be the undisputed leader in
customer satisfaction.”

6
1988 1989
Self assessment Application/site visit
- the value of ‘writing it down’ - taking an enterprise-wide view
- transforming data into information - lack of trend data identified

1991 1990
AS/400 MDQ Gold Award Malcolm Baldrige National Quality
- tying management/measurement Award winner
system to goals - value of teams recognised
- constancy of purpose - necessity of committment to leadership

1992 1993
ISO 9001 site registration Worldwide AS/400 MDQ Gold Award
- continuous improvement process - value of worldwide view
with category champions in place - alignment of business through the
AS/400 quality process

1995 1994
Self assessment Improvement
- value of disciplined decision making - continuous improvement process
- teaming beyod AS/.400 with senior executive ownership
- improved customer focus of five key opportunities

1996 1997
Self assessment Self assessment
- improved key business and - reenergise focus on customer care
customer measurements - expedite skill building in the
- strong customer focus marketplace
- balanced scorecard of measurements
- ISO 14001 registration
A twenty year journey
IBM Rochester’s excellence timelines
1998
Assessment integration
- continual assessments via
internal business processes
20
00

n
Visio Reforming

Strategic
19

Process
80

Process and
product
simplification
Product ISO 9001 and
Enterprise 14001
Market-driven integration Process
quality, Global reengineering
Benchmarking business Business and
Product Customer Empowerment customer-
Zero Manufacturing introduction relationship Baldrige focussed
defects cycle time cycle time management assessment measurements

Leadership

7
According to the 1990 summary, Rochester's quality journey began in 1981 with the
People Responsibly Involved in Developing Excellence (PRIDE) initiative, which
focused on improving product reliability. In 1984 the quality journey was expanded to
include process efficiency and effectiveness, and manufacturing cycle time
improvements. In 1986 the quality journey continued with planning and development
cycle process improvements and integrating suppliers and customers into
development and production processes. In 1989 Rochester's quality journey led to
embracing market-driven customer satisfaction goals, focusing on total cycle time to
the market, and extending customer involvement even further.

Market-driven quality cycle

The market-driven quality cycle of 1990 integrated the six critical success factors into
a closed-loop process..

Strategy
Support customer success
Meet customer needs
Expand the market Requirements

Establish accurate
Gain product and service
satisfied customers specifications

Key
Cycle time Six sigma
reduction
measurements

Create faster
Prevent development,
solutions and
manufacturing and
increase competitive
service rework
advantages

Improve Employee
Employee
productivity education
involvement
and cost

Source IBM Rochester’s Malcolm Baldrige National Quality Award

summary publication, second edition, IBM, Rochester, March 1991

Working toward ‘Six-Sigma quality goals,’ IBM Rochester had reduced development,
manufacturing, and service rework by ensuring that processes produced correct
results the first time. Improving education and employee involvement had increased
productivity and reduced cycle times, enabling faster market response and
competitive advantage. Key measurements verified that processes were definable,
repeatable, and predictable. As this cycle continued, Rochester's strategy was to
‘delight its customers by providing solutions that helped them achieve success in their
business.’

Rochester’s six ‘Critical Success Factors,’ to “ensure continuous improvement toward

Rochester's customer satisfaction goal” were:
1. Enhance total product strategy and plans
2. Improve the requirements definition process
3. Implement Six-Sigma defect-level quality strategy
4. Create and deploy an Excellence in Education plan
5. Enhance and enable employee involvement
6. Develop and implement reductions in total cycle time

8
 Market analysis
Interrelationships between company policy, business goals,
Corporate goals
and segmentation policy, and business and functional strategies and operating plans
instruction

IBM US goals IBM

Competition and world
industry Business and quality
trade
Strategy
Technology ABS goals Executive review Commit
and approval (RMC) operating plan
Business and quality Customer strategies

Supplier
Benchmarks
strategies

Five year strategy - Rochester Operating plan with two-year horizon

Rochester
Business strategy Functional strategy Quality priorities Department, team and
Two year targets
Product; application solution; from strategy employee assessment
Manufacturing - Expand customer
distribution channels; - Revenue of requirements to identify
Development partnerships
support structure - Market - Quality improvement
Market, service and - Improve product and participation - Resource needs
support service reliability - Profit - Training needs
Revenue goals - Reduce total cycle time
Market participation goals - Return on - Tools
Human resources - Deploy quality ownership
assets - Benchmarks
Capital resources
- Supplier improvement
Expenses
Profit goals
Return on asset goals

Market-driven quality, the customer view

Market-driven quality was defined in 1990 to begin with an understanding of what
was important to customers, considering all aspects of their relationship with IBM and
its business partners, from products to support and after-sale service. Satisfying
customers required the ability to provide superior and reliable solutions;
administrative excellence; marketing and sales support; and delivery, maintenance,
and service performance.

Throughout its total enterprise − marketing, product development, manufacturing, and

service − IBM's objective was to provide total customer solutions, demonstrating a
knowledge of the customer's business and all other elements important to customer
satisfaction.

Elements important to customer satisfaction

Administration Market and sales Technical solutions Maintenance and

Delivery
offerings service support
- terms and conditions - low price and high - complete shipments
performance - single person to
(no hidden costs) - shipped on time
- expectations defined contact
- order accuracy - quality and reliability - order-to-invoice time
and set - problem source
- billing accuracy - low cost of ownership shortened
- wants and needs identified quickly
- late bilings - solutions available
understood - knowledgeable
- complaint - incorporate wants and Upgrade
- solutions and customer engineer
management needs into accurate - reduced complexity
applications provided and systems engineer
- telephone specifications - reduced time
- solution affordable - excellent hardware
support - easy to use - flexible policy
- products announced and software service
- reduced cycle time
on time (rapid product introduction) - parts available
- trial period permitted Instability
- rich in functions free of defects
- channel management - short time
- easily migratable sufficient quality
- sales and volumes - not complex near customer’s
- connectivity with - free of defects
increased location
vendors
- customer education - easy to use
- gain and loss management documentation

Marketing branch office Product development Marketing Service

Customer solution
Source IBM Rochester’s Malcolm Baldrige National Quality Award
summary publication, second edition, IBM, Rochester, March 1991

9
The illustration above (previous page) is redrafted from the 1990 application. Below is
the 1999 version, from an IBM Rochester view graph.

Customer view model after revalidation in 1993

Customer satisfaction measurements
Overall Loyalty Recommend
Common attributes
Ease-of-doing-business-with
Partnership
Responsive
Knowledge of customer’s business
Customer driven

Technical Maintenance Marketing Administration Delivery Image

solutions and service support sales offerings
- quality - single point of - information - purchasing - on time - corporate
- reliability contact - solutions procedure - without defects citizen
- availability - flexible - central point - billing - to specification - community
- pricing - available of contact procedure - accurate interests
- ease of use - accessible - education - T and Cs - post delivery - social concerns
- documentation - empowered - empowered - warranty process - environmental
- openess - product - product expiration conciousness
- growth knowledge knowledge notification - technology
- new technology - consistent - competent - financial leeder
- installation/ - ethical alternatives - financial
upgrade stability
- warranty - executive image
- disaster recovery - empathy
- process

Product development Service Marketing Manufacturing Executives

branch office leadership
Source
IBM Rochester Customer solution
Customer view of quality service characteristics

Strategic planning
In 1990 an annually written strategic plan looked forward five years, using
information from market analysis, competitive and industry data, and evolving
technology, in the context of long-term business and quality goals set by IBM
Corporate HQ.
• Customers provided product direction
• suppliers assisted in product sourcing
• teams of employees developed a business strategy that included product and
business goals
• manufacturing, development, marketing, service, and support teams developed
functional strategies that contained the quality improvement plans required to
achieve the strategy.

Functional strategies described the resources, capital, and expenses required to

achieve the quality priorities and the business strategy. Resources and associated
expenses were compared to the revenue and profit goals in order to balance the
strategic plan. Once this strategic plan was approved, an annual operating plan (a
detailed description of how to implement the strategy) was developed with a two-year
horizon.

The Rochester Management Committee, a cross-functional executive team, reviewed

and approved the strategy and operating plans.

10
 IBM Rochester’s product processes in 1990
Planning Development Manufacturing Marketing
Service
Market analysis High-, low-
Segmentation level design Branch office
Component
Position
Simulation manufacturing
Supplier Customer order
Customer manufacturing
Build hardware Order
requirements
test units Sub-assembly validation
Engineering, manufacturing
product,service, verification,
objectives Installation
manufacturing, planning
System
certification 1,2
Customer manufacturing
validation Software, hardware, Installation
component, unit
integration Shipping
Component
Maintenance
specification Customer Sub-system,
Failure service and
validation system test
analysis support

Service Design verification,

planning test, System verify,
test, manufacturing,
certification 3
Manufacturing,
verification, test,
Manufacturing,
certification 4

Customer
Information flow Major transition

Rochester product processes

In 1990 Rochester's production processes were organized into four phases:
planning, development, manufacturing and marketing and service, “tightly linked to
each other and to the customer.”

Cross-functional teams jointly shaped process strategies and planning, identifying

customer needs and converting them into specifications. The development process
took the specifications through a rigorous series of design stages and verification
activities involving both manufacturing and customer evaluations.

The manufacturing process engaged suppliers and the product development team
through joint decision making activities early in the cycle. The marketing and service
process tailored solutions to customer’s needs and represented the voice of the
customer in defining new requirements.

Rochester’s global market analysis and segmentation process gathered data from
sources in all geographic markets − consultant reports, government demographics,
economic forecasts, university studies, and user group feedback − each segmented
into small, medium, and large enterprises and establishments and then into industry,
sub-industry, and application opportunity market segments.

For example, the distribution industry could be segmented into wholesale, specialty,
and retail; retail into drug stores, auto parts, and hard-goods segments (next page).

11
 Market analysis and segmentation

Geographies

Industries Sub-industries
Mfg Wholesale
etc …
Drug stores
Health Distribution Retail Auto parts
Hard goods …
Specialty
Process Finance

small
medium Application opportunity segments
small enterprises large establishments
large enterprises
medium enterprises

Source - adopted from IBM Rochester’s Malcolm Baldrige National Quality Award
summary publication, second edition, IBM, Rochester, March 1991

Selecting target markets

Segment Segment
attractiveness fit
- opportunity - product fit
- growth - distribution
- competition channels
- executive - service and
vision support
- business - application
goals programmes

Analytical hierarchy

Target markets

To select target markets, Rochester evaluated market segments for their attractiveness
and for product and service fit. Attractiveness and fit comparisons were analyzed
using the analytical hierarchy process, comparing the relative strengths of each
market segment with the ability to deliver solutions to it. From this analysis, target
markets were selected and pursued.

Identifying product quality and service features

Plan adjustment
Quality features
and requirements Long term plan
Items
Product
-E
-A
-F
-E Validation sources
Evaluation - -G
-F
weighed against Customer councils
Applications
resources (skills and Short term plan Satisfaction surveys
-B
money) needed Items Consultants
Service
to achieve these -A - Business partners
-C
requirements -B - Cross comparisons
-G - Industries
Channels -C
-D - Field feedback
-D

IBM Rochester performed detailed analyses of targeted markets to identify customer

quality features and requirements. Tools like Conjoint Analysis were used to evaluate
items within four strategic categories: product, applications, service, and channels,
each ranked for its importance within each category.

12
Product quality features and requirements were weighed against the resources (both
people and expenses) needed to achieve those requirements. To balance the plan, the
items were prioritized into short-term (2 year) and long-term (5 year) projects. This
plan was validated through customer councils and marketplace feedback and
adjusted as required.

Defect prevention

Process improvement Cost of defect Defect removal modelling

Preventing design defects shortens cycle time correction Ensures superior product quality
Customer 92x is delivered to schedule
20
installation Model

Hours taken to remove defect

Defects/k lines of code

15 General
Actual
availability
Defects removed

Design Early
10 support
programme

5
Test Testing
13x
Design
x 0
ESP GA
Inspection Verification
Time
Future
development cycle
Design time Test time
1988 development cycle

1985 development cycle

Cycle times improvement was achieved by removing defects early and by

continuously improving design tools and techniques. Quality was improved and cycle
time shortened by eliminating manual operations and errors in interpreting
specifications.

Historical data were used in the software development process to build a defect
removal profile. With this profile, a model was established to track development and
accurately predict the number of defects to be removed at each step in the process.

Before manufacturing began, products were subjected to a series of verification tests

to confirm product characteristics and the integrity of the development process. These
verification tests served as development checkpoints and were coordinated with the
design of the manufacturing processes. An independent assurance organization
certified readiness at each checkpoint and verified adherence to government
regulations, industry standards, and corporate instructions.

Process management approach

One major reason for IBM Rochester's quality results at the 1990 stage of their quality
journey, and the key activity identified at that time for future success was process
management. Process owners defined their processes and identified their customers
and suppliers, aiming to continuously improve quality and reduce cycle times.

Managing processes effectively had resulted in significant cycle time improvements

for computer systems and hard disk development and manufacturing. The serial
processes for producing the System/36 and the System/38 had been transformed to a
parallel, continuous flow system for producing the AS/400 system. Simulation, Early
Manufacturing Involvement (EMI), software component development, and parallel
system tests contributed to a 40% reduction in development cycle time.

13
 Cycle time benefits realised through hardware
simulation and a new software development process
System/36 and System/38 sequential development
Hardware pass 1
Hardware pass 2
Hardware test

Software development
System test

AS/400 synchronised/parallel Ship

development
EVE simulation

Hardware
development with
EMI

Hardware test

10 months saved
Software
prototyping

Software development

Customer review

System test
16 months saved
Ship

Using an ‘Early Verification Engine (EVE)’ to simulate the system design resulted in
80% improvements in the time required to debug hardware from the initial design,
eliminating a need for multiple design iterations.

Early Manufacturing Involvement teams worked out details of product introduction,

software installation, and distribution logistics, beginning in early design stages to
improve the manufacturing processes, enabling rapid, high-volume production
capability. Manufacturing teams build the machines used throughout the development
cycle. Suppliers also participated on EMI teams.

Customers played a major role in clarifying requirements and ensuring that products
met their wants and needs. An iterative software development process was used to
involve customers. Customers were asked to validate the development decisions.
Required changes were made in subsequent iterations.

Continuing customer involvement

To ensure that market and customer needs were accounted for, customers and
business partners were involved throughout the product cycle, from planning through
support and feedback.

14
 IBM Rochester
Customers are involved throughout the product cycle
Customer User groups Software Customer Early Customer
and partner usability availability partnership
business COMMON laboratories testing programme calls and
partner surveys
councils

Analysis Development Build Support

Planning Requirements
and design and test and feedback

Measurement and validation

Innovative techniques established by IBM Rochester integrated customers into

the development process. For example, customer and business partner councils
brought worldwide customers together to review future product plans at an executive
briefing center at IBM Rochester. An independent user group, COMMON, comprised
of over 6,000 customers worldwide, met regularly.

A Software Partner Lab provided an opportunity for customers and business partners
to jointly develop solutions for future product releases. Customers validated that their
requirements were being met, verified that their application programs would operate
effectively on newly designed systems. Customers participated regularly in the
usability testing center to verify that Rochester's products met their operational
requirements.

An early availability program provided AS/400 systems to select customers before

they were available to the general public. These systems were monitored closely to
ensure that customers were satisfied with the latest development-level products,
services, and order and delivery systems. Customer support procedures were also
evaluated during this program.

Contacted by IBM 90 days after receiving an AS/400 system, customers were thanked
for their business and asked about their satisfaction with their new system. See the __
item for detail of how this process works in 1999.

Information and analysis

IBM Rochester had comprehensive, world-class, worldwide information and analysis
systems available, as needed, to all employees in 1990. Information systems
supported communication among employees and with customers throughout the
entire product cycle.

Online tools
An ‘electronic office system’ linked employees worldwide, allowing ‘real-time
communication.’ Online systems ensured that accurate and consistent data was
available for quality improvement and root cause analysis. Rapid communication with
sales representatives, systems engineers and field engineers aided in problem solving
and in communicating opportunities for improvement.

Manufacturing process
The process of moving a product from development to manufacturing was highly
automated. Design data was translated into online manufacturing controls to improve
the quality of manufactured parts. The central control point of the manufacturing
processes was the Manufacturing Control System. Customer orders were
automatically transmitted from the branch office to the Manufacturing Control System
and fed directly to the assembly line.

15
Production was controlled by the customer order, with design data automatically
translated into manufacturing process control information. Assembly and sub-
assembly areas provided process information to manufacturing control, and data were
analysed and reviewed by cross-functional teams for future improvements. Online
systems provided timely and consistent design information to world-wide
manufacturing locations. The automatic translation of information eliminated
opportunities for errors and reduced process steps and time.

Process management
Root cause analysis
Confirm
Recreate problem
Repeat failure
Restate process deficiency
Complete process review
Visit supplier or customer location

Probe
Examine with electron microscope
Examine checkpoints, dumps, and traces
Examine vital product data
Find last point before failure

Analyse
Determine frequency of occurrence (using a Pareto diagram)
Review trend data (trend diagram)
Review statistical process control charts
Decide how exhibited and how found
Evaluate single supplier vs multiple supplier
Analyse execution path, system state, and environment

Correlate
Correlate supplier data, in-house data and customer data
Find indications of problems in other measures
Identify recent process changes
Determine batch, job lot, time stamp and release level

Isolate
Isolate possible causes (using a fishbone diagram)
Rank by probability
Perform additional testing
Perform experiment at suspected point of origin
Test for stress (heat, voltage, strength, life etc)
Test boundary conditions

Online systems provided the data used in root cause analysis − based on an ability to
trace information back to its source. For example, information from manufacturing
control was analyzed to assess changes in new designs or to track problems back to
source. IBM Rochester followed a disciplined approach to track down the causes of
process ‘upsets,’ consisting of the steps in the illustration above.

“This method uncovers the root cause of a problem, identifies required improvements
for products or processes, and expands existing knowledge of the complex
interrelationships involved in satisfying customer needs," the 1990 summary
concluded.

Rochester improvement teams

Process improvement teams were used extensively to simplify processes, create new
defect prevention methods, and reduce cycle times in product and non-product
processes.

For example, the transformation of the engineering change process from traditional
printed copy to an online environment was successfully completed by a quality
improvement team. The process owner, a line manager, was also a member of the
team. Working closely with its internal customers, the team began by modeling the
process, using root cause analysis to determine what caused process upsets.
16
The process was restructured and unnecessary steps were removed. To implement
the new process, education and support were provided to customers across the site.
Business gains were realized in product cycle time, efficiency, and communication.

Statistical process control

Statistical Process Control (SPC) was used in critical operations throughout the
Rochester site and its supplier’s sites. EMI teams worked with development engineers
to identify critical parts, parameters, or process steps that require continuous control
through statistical monitoring techniques.

Continuous flow manufacturing

IBM Rochester had a process called continuous flow manufacturing integrating all
elements of its production and non-production support processes to achieve:
• reduced cycle time
• reduced cost
• continuous focus on defect prevention

Continuous Flow Manufacturing begins with an examination of the total process, from
a customer order to raw material gathering through manufacturing and customer
installation. Order sizes, improvements in tooling and changes in process flow are
some of the factors affecting CFM cycle time.

IBM Rochester AS/400 order process

Expert system
configuration Plant and manufacturing
control system
Order entry
and validation Build order and assembly
Customer Branch office sequence control
order
and marketing

Distribution

On-line ordering was in use at Rochester in 1990, linking customers to the AS/400
production process. Orders, configured with the assistance of sales representatives
and validated when entered into the system, were routed directly to manufacturing
control, where build order and assembly sequence information was generated and
sent to the plant floor.

Human resources
Achieving market-driven quality goals depended on enabled, empowered, excited
and rewarded employees, the 1990 application summary said.. IBM Rochester’s HR
strategy attempted to bring about a cultural shift from a product-driven focus to a
market and customer-driven quality focus, and consisted of three initiatives:
1. formal education
2. on-the-job customer contact
3. participation.

Morale, buy-in, participation and productivity were measured to establish baselines

and drive continuous improvement

17
Integrating people into quality and business plans

Education and communication

IBM Rochester is continually training employees and developing
their skills - spending on education is five times the national
average. Individual discussions between employees and managers
is the most common and important communication channel.
Basic belief: respect for the individual Other avenues include:
Education partnership Education and training Management development Communications
Site goal - faculty loans - quality education - quality training - employee-manager
cooperative education - manufacturing skills - management college discussions
Business plans People vision programmes integration - shadow programme - regular department
- enable - tech interchanges - voluntary education - mentor programme meetings
- empower with universities - technical vitality - tech assisitance - functional meetings
- excite - Greater Rochester - graduate work study assignments - IBM TV network
- reward University Centre - supplier education - cross-functiona; - bulletin board notices
assignments

Human resource strategy

Customer contact
Paradigm shift to market-driven culture
All employees from senior managers to production staff
Awareness understanding attitude behaviour
are given opportunities to have regular contact
with customers
People initiatives
On the job market-driven experience
- advocates
- on-the-job customer contact
- telemarketing assignments
- participation
- customer partnership call assignments
- executive assistance to marketing
- key account owner
Measures of effectiveness
- installation assessment team
- morale
- productivity
- buy-in
- paticipation Participation
IBM Rochester fosters staff participation and empowerment.
Compensation and recognition programmes ensure that
people are rewarded for improvement and innovation

Participation initiatives Decision-making Job flexibility Recognition Compensation

1,200
- IBM Rochester continually - opinion survey - full employment - performance - merit pay
increases participation in - ‘Speak Up’ process - volunteer temp evaluation - IBM benefits plan
technical vitality initiatives, 800 - open door programme assignments - vebal thanks - voluntary tax-deferred
measured here by employee - Suggestion plan - flextime scgedules - promotion savings plan
participation in catagories 400 - roundtable discussions - internal transfer - ABS Quality Awards - vountary stock-purchase
such as writing articles in - skip-level interviews - retraining - management excellence option
professional journals, patent - development plans awards - employee assistance
0
applications and inventions. 1986 1987 1988 1989 - performance plans management appreciation programme
- manunfacturing skills awards - ‘Plan for Life’
integration - author recognition - Elder care referral
- inventor recognition - child care referral
- site-wide publicity - IBM club
- Watson Scholarship

Customer satisfaction processes

Customer partnership call

- To say thanks for purchasing
- To ask about likes and
dislikes
Summarises call Establishes Tracks
- to seek comments
Customer information marketing dissatisfaction
team contact plans to
record closure
Database record-keeping
- data
- detailed comments Analyses for Contacts local 30 day
- analysis of responses trends and branch manager closure
improvements process

Monthly summary Makes customer

report for contact
Customer No
improvements
satisfied? given to
manufacturing Action plan
Yes development, to resolve
marketing and dissatisfaction
service within 30 days
Satisfied
customer

The Customer Partnership Call process thanked customers for purchasing an AS/400.
The call was (and still is, see the Customer calling at IBM Rochester exhibit) made 90
days after shipment from Rochester, seeking the customer's likes and dislikes, as well
as any comments they may have. These comments were placed in a database, ana-
lyzed, and distributed regularly to engineering, programming, marketing,
manufacturing, and service teams for evaluation.
18
Dissatisfied customers or customers with concerns were contacted by marketing to
understand their concerns in more detail. The Customer Satisfaction Project Office
received notification of the results of the customer contact. Thirty days later, the same
customers were called back to ensure they were satisfied. Other methods to ensure
leadership in customer satisfaction included customer satisfaction surveys,
independent consultant reviews, and industry-accepted reports.

Complaint management processes

at IBM Rochester
Customer Customer
complaint feedback programmes

Customer satisfaction
Corporate/LOB
management team
- sends feedback
complaints
to branch office
Area/branch - records complaints
- analyses complaints
Investigator
assigned
Analysis of results
the same day
presented to:
- customer satisfaction
Close complaint council
within 2 weeks - customer satisfaction
process council
- LOB general manager

Product, service,
and process improvement

Measure and validate

Source - IBM Rochester 1991

To achieve its goal of undisputed leader in customer satisfaction, IBM Rochester

established a complaint management process that tightly linked marketing and
service teams to development and manufacturing teams.

Complaints were managed by a Customer Satisfaction Management Team working

with the branch office closest to that customer. An investigator was assigned to
investigate and understand the specific details of the complaint, which had to be
resolved within two weeks. Feedback from the complaint is sent to the Customer
Satisfaction Management Team where it is recorded, assimilated with other
information, and correlated for use in reports to the Customer Satisfaction Council,
resulting in product and service quality improvements.

19